Hi--
>Bob:
>Thanks for the email. I gather that you have done some objective tests. I
>appreciate your info.
--If you look up the Subjective Lens Evaluations, you know that I have checked
--a LOT of lenses, though I did not cover the non-Nikkors that do not fit Nikon
--bodies.
>> In my experience and in that of friends, variability is a BIG issue.
>> Nikon lens manufacturing is pretty consistent
>< snip >
>> ... 35-105mm stands out - only 2 in 10 that I tried were well-aligned).
--And only 3 of 10 were sharp at edges and corners.
>> Other lens lines show MUCH greater variation, even a scary level of
>> variation (to the point that I would not consider buying lenses in
>> those lines without testing and selecting every lens).
>
>Wow! That sounds scary! I'm not surprised that Nikon stands up well. How
>does one test alignment? Is this something I can do at home, or do I need
>a bench & optical equipment.
>( My 100-300 zoom got banged a few times, I worry about it, because it
>seems so flimsy.)
--I do not believe in formal tests. there are some simple tests you can do
--that are fairly reliable (see below).
> > > >Does anybody have any data on this issue? Should consumers "shop
>> >around" and test several different copies of the same lens to choose
>> >the sharpest one?
>
>> Yes! If you care about lens quality - BUT, beware of developing an
>> obsession! (When you begin to realize just how high a percentage of
>
>< snip >
>
>Ok. My next purchases are likely to be a 20-35 f3.5 USM and then a 300
>f4. I doubt if the local shops have a bunch of these in stock, maybe a
>few at a time.
--One advantage (besides price) of buying from B&H is that they will keep --feeding you replacement lenses until you are happy. I am finicky - it
--took 5 105mm f1.8's to find one I liked (BUT, any of the ones I rejected
--would have satisfied anyone else - all were sharp, but 4 showed a minor
--misalignment detectable wide-open). It only took one 35mm f1.4, even though
--that is a much more complex design and includes floating elements.
> How do I approach testing them? Do I just buy one, take it
>home, test it & bring it back if it's not tack sharp? Or do I mount 2 in
>the store, shoot slides from a tripod, & buy what appears to be the
>better of the two?
--Either is good, if the store is willing - ask first (and check on stock).
>My problem is not knowing what to expect. Like, I expect that my 100-300
>f4.5-5.6 at 300mm will not blow up to 11x14 as sharp as my 100 2.8 macro
>(my slides verify this too), but how do I know just how much to expect
>from the 100-300 before saying that its either OK or a dog?
--(See below about standards.) A quick check for good alignment is the --following: With the aperture set at f2.8 and f5.6 for wides and zooms,
--f2 and f2.8 for normals, and wide-open for teles, carefully focus on a
--distant, detailed subject (with the detail running along the top edge
--of a vertical frame with the camera held with the grip up. Take a picture,
--then invert the camera carefully, without disturbing focus, and take the --same picture (preferably on B & W film). With zooms, you will need 3-4
--pairs to cover the range - refocus at each focal-length. Shoot straight-down
--at a textured ground one frame for a short-mid distance check at f5.6. Shoot --at minimum focus of a newspaper, also at f5.6 (Alignment with subject is
--difficult). (Repeat checks that show anomolies - it could easily be technique)
--(it gets better with practice). With a GOOD 10x magnifier (can be cheap - $5 --magnifiers used for scientific field exams work well), check the negatives.
--With the distant pairs, you have the same details across the frame-line.
--You should see good detail to the corners and equal detail level in all four
--corners. The two edges should look identical in detail level, and at the --center of the edges, you will have the same subject in both frames - it should --look identical in both frames. The ground shot should show equal corners, and --equal edges (ideally, the edges and corners should match the center in --quality, but if the corners are still decent, accept it). The close check
--is similar (use the newspaper columns and edges to help you align the camera
--with the subject). Here, many otherwise fine lenses fail to be very sharp at --the edges and especially the corners - if they are well-aligned, accept them,
--even if a bit soft (see notes in "SLE(MN)".
>I resort to the Pop Photo guides to form opinions about lenses, bcause
>they are quite detailed. Their favorable review is one reason why I want
>to get a 20-35, and am not rigidly sticking to a 24 mm prime. However -
>this is between lens comparison, not within the same lens model.
--Pop's test, unfortunately, do not say much about differences in distance
--performance (some lenses vary a lot in performance at different distances).
--The tests also are a bit vague about edge-corner performance - to me, a lens
--that is not sharp in the corners at most distances is not sharp. No lens is --perfect, so one must establish one's own standards of acceptable quality.
--For me, normals must be good by f2, wides and zooms by f5.6, teles wide-open --(center to --corner) (they are often better), and they MUST show no more than --minimal misalignment. Misalignment is the give-away of manufacturing problems, --and it will usually show up before serious sharpness degradation is evident.
> >
>> There may be some association, or it may be incidental, but the Nikon
>> lens line is also remarkably consistent in the quality of lens design
>> throughout the line regardless of price (lens barrel quality is another
>> matter....)
>
>I am sure that some brands (e.g. Nikon) are pretty consistent. I expect
>that if optical quality is good for lens A it is likely to be good for
>lens B. The glass is probably the easy part.
--Bad assumption....
>Now - without getting into a brand war, what about Canon - for same-lens
>to same-lens consistency? I assume that they are *in-general* quite good.
>I use Canon. I would not argue the Canon-is-better line, but I do have a
>bunch of money invested and will be unlikely to change equipment. On the
>other hand, if you have found that Canon is a dog - I'd start to worry. I
>appreciate your opinion.
--Unfortunately, both Leitz and Canon are not great for consistency in
--either design or manufacture - I have seen a lot of defects and bad designs
--in both lines (didn't dare mention this in the post - just the Japan-Germany
--comment). If you are willing to search for well-aligned samples of the
--better designs, Leitz and Canon are fine. Hope this helps.
--David Ruether